When you come up with a new idea and develop a new trading strategy based on it, then you call up historical data for back-testing and get a result. You feel unsatisfied, so you adjust a parameter, such as the 5-day moving average. Changed to the 10-day moving average, tested again, and found that the result was slightly improved. You adjust another parameter, and this process continues. Some adjustments improved the results, some worsened the results, and finally, you finally called up a set of perfect parameters that made the results of the strategy surprisingly good.
Congratulations, you have overfitted! This kind of operation is a classic mistake often made by typical amateur quantitative researchers.
If your strategy has good test results for this series of tuning parameters, then your strategy is likely to find the real law, which is not overfitting.